[NTLUG:Discuss] Routing Problems?

[Jason Smedley]

I've got a very strange problem with one of my Linux RedHat 9 severs.
 
This Linux server will, after a period of time, loose the ability to
ping two devices on a remote network connected by a point to point T1.
The two devices in question are periodically pinged to verify up/down
status.  I believe that if it was communicating with other devices on
the remote network, that it would also loose communication with them.
The time between communication loss can vary anywhere from a couple of
weeks to a few minutes.
 
I cannot ping the remote router, or a particular host, I can however
telnet to the remote router from the Linux box so it seems that it
doesn't loose complete communication.  Rebooting the Linux server fixes
the problem for a while, and then it looses communication again.  It has
no problem communicating with other devices on the LAN, or other devices
on the remote network.  Other devices on the same LAN as the Linux
server have no problem communicating with the remote router or host.  
 
  I can also temporarily fix the problem by putting in a static route on
the Linux box to the remote network or host, but it eventually looses
the ability to communicate again.  Removing the manually added route
also re-establishes communication, for a while.  It seems like the act
of doing changing the Linux routing tables temporarily fixes the
problem.
 
If I perform a traceroute to the problem host, it doesn't even get to
the default gateway.  It acts as if it's trying to reach a host on the
LAN.
 
The Linux server's default gateway points to the local router of the
point to point T1.  
 
I'm sure I didn't explain it very well, so I'll be glad to clarify
anything as best I can.
 
Any ideas, suggestions?
 
Thanks